]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
overview / pkg / ggml / sources / llama.cpp / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: d00cbea)
cpu : optimize the ggml NORM operation (#15953)
authorduduta <redacted>
Thu, 9 Oct 2025 19:11:15 +0000 (22:11 +0300)
committerGitHub <redacted>
Thu, 9 Oct 2025 19:11:15 +0000 (21:11 +0200)
commit1deee0f8d494981c32597dca8b5f8696d399b0f2
tree2db38ad160ae80c043fd1dd2bfd98d6c91945e96tree
parentd00cbea63c671cd85a57adaa50abf60b3b87d86fcommit | diff
cpu : optimize the ggml NORM operation (#15953)

* ggml-cpu: optimize norm operation to use intrinsics or Accelerate

          rename function

          add endif macro comment

Co-authored-by: Georgi Gerganov <redacted>
Co-authored-by: Aaron Teo <redacted>
* implement s390x SIMD suggested by @taronaeo

* add TODO comment

* tidy up spaces

---------

Co-authored-by: Georgi Gerganov <redacted>
Co-authored-by: Aaron Teo <redacted>
ggml/src/ggml-cpu/ops.cpp diff | blob | history
ggml/src/ggml-cpu/vec.cpp diff | blob | history
ggml/src/ggml-cpu/vec.h diff | blob | history
Packaging of ggml-org/llama.cpp